A stochastic model for the large scale dynamics of some fluctuating interfaces (Q1118257)

The Kuramoto-Sivashinsky-Tsuzuki equation [see, \textit{Y. Kuramoto} and \textit{T. Tsuzuki}, Prog. Theor. Phys. 55, pp. 356 (1976); \textit{G. I. Sivashinsky}, Acta Astronaut. 6, 569-591 (1979; Zbl 0398.76070), and references therein] and a similar model describing irregularly modulated interfaces are investigated in a very large length regime. This corresponds to a permanent chaotic state with a large number of nonlinear degrees of freedom. The dynamics of the small wavenumber modes is analyzed. A simulation of the equation allows to show that these very slow modes obey an effective stochastic partial differential equation, i.e. a Burgers equation with a rapidly fluctuating forcing term. This deterministic forcing ``noise'' is computed numerically. A surprisingly robust property of this noise is that its autocorrelations decay rapidly in time for all parameter values and wavenumbers. A consequence of the existence of such an effective equation is that the correlations of the slow modes should scale as if they obeyed a Burgers equation with a stochastic forcing. For this latter equation predictions made using the dynamic renormalization group and direct simulations exist. These predictions are compared to the data obtained for the spontaneously fluctuating equations. The method employed in this paper to demonstrate that the slow modes obey an effective stochastic equation is of some generality and can be applied to other chaotic partial differential equations.